110 likes | 253 Vues
Paris, August 2005. IETF 63 rd – netlmm BOF. Requirements and Gap Analysis for IP Local Mobility (draft-kempf-netlmm-nohost-req-00). Gerardo Giaretta James Kempf Phil Roberts Kent Leung Katsutoshi Nishida Marco Liebsch. Purpose of the document. The I-D has two main objectives
E N D
Paris, August 2005 IETF 63rd – netlmm BOF Requirements and Gap Analysis for IP Local Mobility (draft-kempf-netlmm-nohost-req-00) Gerardo Giaretta James Kempf Phil Roberts Kent Leung Katsutoshi Nishida Marco Liebsch
Purpose of the document • The I-D has two main objectives • list the requirements for a NETLMM solution • highlight the gaps between state-of the art solutions and requirements • Solutions identified so far • Mobile IPv6 with local HA assignment • Hierarchical Mobile IPv6 (HMIPv6) • Combinations of Mobile IPv6 with optimizations • MIPv6 with local HA assignment + FMIPv6 • HMIPv6 + FMIPv6 • Micromobility Protocols (e.g. Cellular IP, HAWAII)
NETLMM Requirements • Req #1 – Handover performance improvement • handover packet loss and handover latency must be minimized • this is to fulfill requirements of real-time applications on jitter, delay and packet loss • Req #2 – Reduction in handover-related signaling volume • signaling volume to handle handover should be minimized • e.g. movement detection signaling, location update signaling, security-related signaling
NETLMM Requirements (cont’d) • Req #3 – Location privacy • location informationshould not be revealed to nor deduced by the correspondent node without the authorization of the MN • draft-koodli-mip6-location-privacy-01 • Req #4 – Efficient use of wireless resources • minimization of per packet overhead over the air interface • state of the art so LMM solutions increase packet size over the air by adding tunneling or other per packet overhead • header compression can remove header overhead but it increases the cost and complexity of the access points (i.e. higher per packet processing across the wireless link)
NETLMM Requirements (cont’d) • Req #5 – Reduction of signaling overhead in the network • signaling within the wired network should be minimized • this is mainly for reducing cost of laying fiber or wiring to the wireless access points in a widely dispersed geographic area
NETLMM Requirements (cont’d) • Req #6 – No extra security between mobile node and network • LMM protocols involving signaling between host and network require additional SAs between the host and one or more network entities • establishing a SA specifically for LMM may require extra signaling • establishing a SA specifically for LMM may be difficultin a roaming scenario (i.e. potential barrier ro deployment) • removing host involvement also limits the possibility of DoS attacks on network infrastructure elements
NETLMM Requirements (cont’d) • Req #7 – Support for heterogeneous wireless link technologies • handover between different wireless link technologies • Req #8 – Support for unmodified hosts • no host software installationon the user terminal • extremely successful in the WLAN switching market • enables a service provider to offer service to all customers • multiple global mobility management protocols can be supported • Req #9 – Support for IPv4 and IPv6
Main gaps for current solutions • Mobile IPv6 with local HA (+ FMIPv6) • FMIPv6 is needed (req #1) • high signaling volume if route optimization is used (req #2) • FMIPv6 requires additional signaling • location privacy only with bi-directional tunneling (req #3) • if no RO, over-the-air tunnel to the HA (req #4) • further temporary level of tunneling between MN and PAR in FMIPv6 • bootstrapping a SA with a local HA is needed (req #6) • FMIPv6 requires an additional SA with the ARs • host support for MIPv6 and FMIPv6 (req #8) • IPv4 support needed for both MIPv6 and FMIPv6 (req #9) • miptrans DT in mip6/nemo WGs
Main gaps for current solutions (cont’d) • HMIPv6 + FMIPv6 • FMIPv6 is needed since HMIPv6 only partially shortens handover latency (req #1) • HMIPv6 reduces handover related signaling volume since no RO signaling is done for intra-MAP handovers (req #2) • FMIPv6 still requires additional signaling • tunneling between MN and MAP (req #4) • further temporary level of tunneling between MN and PAR in FMIPv6 • SAs needed between MN and MAP (for HMIPv6) and MN and AR (for FMIPv6) (req #6) • host support for HMIPv6 and FMIPv6 (req #8) • IPv4 support needed for both HMIPv6 and FMIPv6 (req #9)
Main gaps for current solutions (cont’d) • Micromobility protocols • host route propagation is required throughout the wired network (req #5) • most of the requirements are fulfilled • potential drawbacks from a deployment and scalability standpoint • involve every routing element between the MN and the LMM domain boundary router in all packet forwarding decisions • scalability is limited because each care of address corresponding to a MN generates a routing table entry